gtkradiant

History

rambetter f487ea7c54 Updating notes on regression tests. They are "mostly sort of fixed". Won't be fixed for good until I look at the last bit of code that has not been examined yet, which is the plane intersection code. I want the errors to be much less than they are now, even though the disappearing_sliver* tests are now working. git-svn-id: svn://svn.icculus.org/gtkradiant/GtkRadiant/trunk@378 8a3a26a2-13c4-0310-b231-cf6edde360e5		2010-12-29 05:20:32 +00:00
..
maps	Adding sparkly_seam q3map2 regression test. No fix yet, but there will be	2010-12-27 23:14:32 +00:00
textures/radiant_regression_tests	Adding sparkly_seam q3map2 regression test. No fix yet, but there will be	2010-12-27 23:14:32 +00:00
README.txt	Updating notes on regression tests. They are "mostly sort of fixed". Won't	2010-12-29 05:20:32 +00:00

README.txt

DESCRIPTION OF PROBLEM:
=======================

The example map, maps/sparkly_seam.map, contains two triangular brushes near
the side of the room.  The seam between these two brushes "sparkles" even
though the endpoints of the edges are exactly the same.

To trigger the bug, compile the map; you don't need -vis or -light.  Only
-bsp (the first q3map2 stage) is necessary to trigger the bug.  The only
entities in the map are a light and a info_player_deathmatch, so the map will
compile for any Q3 mod.


SOLUTION TO PROBLEM:
====================

None yet.  The problem is likely caused by sloppy math operations (significant
loss of precision).  This bug pops in and out of existence with every other
commit at the moment.  The problem is likely caused by the operations in the
brush winding computation (where the planes are intersected with each other).
I have not gotten around to addressing that code yet.